BA Prince the B.A. Subscribe

Home / Jira & Confluence for BAs

Spikes in Jira: How to Create a Timeboxed Research Story for Unknowns

Spikes in Jira: How to Create a Timeboxed Research Story for Unknowns

Watch on YouTube Subscribe

When the team cannot estimate a story because nobody knows the answer yet, write a spike: a short, timeboxed piece of research. Create one in Jira with a clear question, a timebox and the output you expect.

Step by step

What a spike is

  1. Spikes in Jira. When the team can't size a story because nobody knows the answer yet, write a spike. A spike is a short, timeboxed piece of research. It produces knowledge, not code. Here we create one in Jira.
  2. Create. This is the backlog of an online shop project. Click Create at the top of the screen.
  3. Pick Task. Jira has no Spike type by default, so most teams use a Task. Click the type field at the top of the create window. Pick Task, the one with the blue tick icon.

Write the spike

  1. Spike in the title. In Summary, type "Spike: Can we show live parcel tracking?" Starting with Spike makes it easy to spot on the board, and the rest states the unknown as a question.
  2. The question. In Description, write the one question the spike must answer, in plain words: "Question: Can the shop show where a parcel is right now, using the courier's tracking data?"
  3. The timebox. Next, add the timebox: "Timebox: 2 days. Stop when the time is up." The research ends after 2 days, even without a perfect answer, so the spike can't take over the sprint.
  4. The output. Last, say what the team hands back: "Output: a short write-up, a yes or no, and new stories with estimates."

Link and create

  1. Parent epic. Link the spike to the epic it supports. Click the Parent field (in Jira Data Center and some older projects this field is called Epic Link). Parcel tracking is part of order tracking, so pick the order tracking epic.
  2. Create. When the fields are filled in, click Create to save the spike.
  3. In the backlog. The spike appears at the bottom of the backlog, ready to plan into a sprint like any other item. Click it to open it.
  4. The timebox. The timebox is in the description, where the whole team can see it. When the 2 days are up, the team shares what it learned and closes the spike.

Recap

  1. Question, timebox, output. A spike is research, not code. Start the summary with Spike, ask one clear question, set a timebox, and say what the team will hand back.

Common mistakes

  • No timebox. Without a fixed limit, research keeps going and eats the sprint. Write the timebox in the description, like "Timebox: 2 days", and stop when it runs out, even if the answer isn't perfect.
  • Several questions in one spike. A spike that tries to answer everything has no clear end. Ask one question per spike, and create a second spike if another unknown comes up.
  • Closing the spike with nothing to show. If the output isn't written down, the team learns something and then forgets it. Agree on the output up front, such as a short write-up, a yes or no, and new stories with estimates.
  • Treating spike code as finished work. Any test code from a spike is there to answer the question, not to ship. Put the real build in the new stories the spike creates, so it gets estimated and tested properly.

When a business analyst uses this

In backlog refinement, the team can't size a parcel tracking story because nobody knows if the courier's tracking data supports it. The BA writes a spike with the question, a 2-day timebox and the expected output, links it to the order tracking epic, and turns the yes or no into new stories for the next refinement session.

Questions people ask

What is a spike in Jira?

A spike is a short, timeboxed piece of research the team does when it can't estimate a story because nobody knows the answer yet. It produces knowledge, like a yes or no and new stories, not finished code.

Is there a spike issue type in Jira?

Not by default. Most teams create a Task and start the summary with "Spike:". A Jira admin can also add a custom Spike type if the team wants one.

How do you write a spike in Jira?

Create a Task with a summary like "Spike: Can we show live parcel tracking?". In the description, write the one question, the timebox and the expected output, then link it to its epic.

More in Jira & Confluence for BAs